@scenerok/cli 2.2.0

package/skills/remote/brand-setup.md

@@ -0,0 +1,69 @@
+# Brand Setup for Agents
+
+Collect brand context **only when it improves the video** and **only when your runtime allows asking the user or discovering assets**. Never block composition on a long questionnaire.
+
+## When to ask (vs infer)
+
+| Situation | Action |
+| --- | --- |
+| User gave product/company URL + you can browse | Extract colors, logo, hero imagery from the site first |
+| User mentioned a brand by name with no assets | Ask for logo URL and primary color (one message, ≤3 fields) |
+| Template has `color` placeholders | Use template defaults; offer to customize |
+| Motion graphic / code demo with no brand | Skip brand questions |
+| User already specified hex colors / logo | Use as-is; do not re-ask |
+
+**Ask at most one short brand message** with combined fields, not a multi-step wizard.
+
+## What to collect (priority order)
+
+1. **Primary brand color** — hex, e.g. `#6366F1` (for text `color:`, lower-third `accent:`, animation `accentColor:`)
+2. **Logo** — HTTPS URL or uploaded asset (editor library / ChatGPT import / local path on CLI)
+3. **Secondary / text color** — default white `#FFFFFF` on dark or dark `#0F172A` on light
+4. **Font preference** — only if user cares; default `Inter` / system stacks in examples
+5. **Tagline or CTA** — for promos and `logo_outro`
+
+## Where brand lands in VidScript
+
+```vidscript
+import motion from "@scenerok/basic-animations"
+import animations from "@scenerok/animations"
+
+input logo = "https://cdn.example.com/brand-logo.png"
+
+let brandColor = "#6366F1"
+
+[0s .. 3s] = text "Launch Week", color: brandColor, style: title, animate: motion.popIn(0.6s)
+[3s .. 8s] = animations.lt_clean_bar("Product Name", "Tagline here", accent: brandColor)
+[8s .. 14s] = animations.logo_outro("Thanks for watching", url: "scenerok.com")
+```
+
+Use `accent`, `accentColor`, and `color` params consistently across text and animation blocks.
+
+## Per-environment collection
+
+| Environment | How to get brand assets |
+| --- | --- |
+| **Web editor** | Ask in chat; user can upload to asset library and paste URLs into script |
+| **CLI / desktop agent** | Ask in terminal; accept local paths; user may run `scenerok auth login` for cloud assets |
+| **ChatGPT** | Ask in conversation; user uploads logo/screenshot → `importOpenAIConversationFiles` → use returned `asset.url` |
+| **URL-only (no uploads)** | Screenshot logo from site; probe dimensions before placement |
+
+## Website-derived brand (preferred when URL given)
+
+1. Visit the URL (browser tool or user-provided screenshots).
+2. Note dominant accent from buttons, links, or hero — sample one hex.
+3. Save logo and product stills; **probe width/height** before composing.
+4. Do not invent claims, pricing, or testimonials — extract or ask.
+
+See `workflow.md` for the full URL asset pipeline.
+
+## Defaults when user declines
+
+- Accent: `#6366F1` (SceneRok indigo) or high-contrast pair white on `#0B1020`
+- No logo: kinetic `text` + `motion.popIn` instead of `logo_outro`
+- Music: neutral instrumental bed from `music.md`
+
+## Privacy
+
+- Do not paste API tokens or OAuth secrets into VidScript or chat logs.
+- Use only assets the user provided or that are public on their stated URL.